Iyfjky

Question

Given the amazing computational power of Matrioshka Brains (M-brains), all the amazing things it can do, including but not limited to the following.

Allotting space for an almost perfect simulation of the continuation of the last natural configuration of the star system. With a formalized Grand Unified Theory (an equation governing all the fundamental physical laws of the universe), M-brains can even be made to think about futures where they didn't exist.

Lifting the parent star to extend its lifespan and/or extract materials that the resident celestial bodies (like asteroids, planets, moons, etc.) cannot provide.

Housing digital souls in different GUT-based simulations.

Engineering each node to be capable of housing, securing, and providing for physical beings as well.

Other than the fact that they may simply choose not to, why wouldn't advanced civilizations litter the galaxy with these monstrosities?

Instead of:

Launching generation ships

Searching for habitable planets

Terraforming uninhabitable ones

Burrowing into asteroids

Why not launch Von Neumann probes that can terraform entire star systems into livable M-brains?

If developing civilizations are still living within the star system, surely an ethical dilemma will be encountered, but its advantages far outweigh the disadvantages so convincing natives to participate and cooperate with the M-brain creation will be the only logical course of action.

What has terraforming to do with matrioska brains? Please check your tags — Aug 7 at 19:20
Perhaps the better question is why would advanced civilizations do these things? At least that's the question you need to answer first. What are the motives of these advanced civilizations? — Aug 7 at 19:56
From what I read, the capabilities of a M-Brain are not simple to comprehend. I don't think it's clear whether it's possible to simulate the star system that you consumed (indeed, I have reasons to believe it is impossible for information theoretic reasons). Are we to assume that your M-brains do have all of the specified capabilities? — Aug 7 at 20:43
Why would we want to turn every star system into a Matrioshka brain? What are we gonna do with that many Matrioshka brains? Simulate the star system? We could simulate the star system just by doing nothing and save a bunch of effort. — Aug 8 at 6:00

score 5 · Accepted Answer · 2018-08-22 12:37:12Z

I can think of a number of reasons why a civilisation that managed to build one Matrioshka Brain wouldn't ever build another, here are some of them:

slow-boat travel issues, the distances between stars means that sending conversion systems to other stars takes millions of years, assuming nothing goes wrong on the trip, a Matrioshka civilisation would of course simulate any action before taking the plunge. Sending out equipment packs to convert the neighbours will take millions of years and has at best a low chance of success. Knowing that their civilisation will never see such a plan come to fruition there's no reason to try.

light speed, so you convert neighbouring systems to create new Matrioshka Brains and? Without some sort of instantaneous communication you still don't have neighbours you can have a meaningful conversation with, there's not much point in expending resources on grand projects that don't actually do anything to enrich your existence.

induced isolationism, having a Matrioshka Brain, you can do anything you want to, as individuals and as a society, without ever leaving home anything. Such a society, with nothing to strive for, will stay home and vegetate, uninterested in anything in the outside universe.

economic collapse, building a Matrioshka Brain may be so expensive of time and resources as to deplete the civilisation that undertakes the project to the point where they simply can't mount another large scale project like an interstellar mission.

physical collapse, a Matrioshka Brain is potentially far more vulnerable to radiation damage, from events large and small, than any other form of civilisation ever thought up. A large event like a Gamma Ray burst, or a series of small ones like a flare swarm, could cripple a Matrioshka civilisation still in the construction phase, before it can look to send out envoys.

anthropic incident, if the universe is at all anthropic then the creation of a Matrioshka Brain is actually extremely dangerous. Since a Matrioshka Brain is super-sapient and super-sentient if the universe is weakly anthropic it could in fact wipe out the civilisation that tried to build it by thinking of a world without itself in it, or by thinking of something worse. If the universe is strongly anthropic a Matrioshka Brain simulation of a universe without it, like the one you propose, would result in said Matrioshka Brain disappearing along with the civilisation it supports.

You may have simply overlooked stuff like: This civilization has a formalized GUT. They can even use it in their simms. They are doing this to have neighbors. Judging by the customized heavens offered by M-brains, they're not just trying to enrich their existence: they're trying to enrich existence itself by making more of themselves exist. Inevitably, at least one of those people will want to be on other real star systems, and she/he can be in for the next project. Only probes will be sent out to start such projects. M-brains are the most vulnerable, but it's also the only standing option. — Aug 15 at 17:31
@KyleZabala There is no reason to believe that people will become any less self-centred and greedy because they make a shiny Matrioshka Brain, they'll wallow in their creation not try to improve the universe, that covers points 3 and 4. As to a Matrioshka Brain being the "only option" that's rubbish there's always another option, even if it's only "go extinct", civilisations that really go the distance will be the ones that choose the absolute safest development pathways, ones that lead away from Matrioshka Brains because of their obvious physical vulnerability. — Aug 15 at 17:53
@Ash I'm not sure you understand the anthropic principle if you believe that a computer running a simulation can cause itself to vanish. Do you want to elaborate more on how your proposed "anthropic incident" works? — Aug 15 at 18:59
@KyleZabala We don't know where Gamma Ray Bursts come from, we think they might come from supernovas but we've never seen one straight on to be able to get an origin fix, and we probably won't survive getting that data. A true Matrioshka Brain is a one trick pony, if you want to starlift, etc... you need other toys that compromise the Matrioshka Brain's functionality. The star at the core of a Matrioshka Brain is a bigger threat to it than any outside star and you don't want to mess with your civilisation's sole power source so there's no proven reason to mess with any star. — Aug 15 at 19:19
@Ash, no, that's not what anthropic means. The anthropic principle is based on the fact that the universe seems improbably perfectly suited to life. That claim suggests intelligent design. The anthropic principle answers that, no, it has to do with conditional probability. These conditions may be unlikely given no information but, given that we (human observers) exist, the probability is 1. It doesn't mean that "the observer" brings the universe into being. That's solipsism or quantum jiggery-pokery of some kind. — Aug 15 at 20:30

AugustDay 2194 · Answer 2 · 2018-08-07 20:18:02Z

Material Limits:
Matrioshka Brains are big. A megacomputer so large it is wrapped around a star in multiple layers. It is virtually impossible to comprehend just how big such a structure would be, if you were actually living in one. As Robert Bradbury puts it in this 1999 paper,

Matrioshka Brain construction is limited by the fundamental abundance of elements in their local region of space. Silicon may be excellent for building microprocessors and carbon (as diamond) excellent for building rod logic computers but neither of these elements is highly abundant in the universe as a whole. A major part of engineering MB will be the efficient partitioning of matter into the various components.

He goes on to describe some specific scenarios that would limit the construction of such a structure based upon the local situation:

Insufficient titanium, aluminum or carbon to build the maximum number of nanoscale components, particularly high pressure circulation systems and nanocomputers.

Insufficient aluminum or magnesium to build solar collecting apparatus capable of harvesting and redirecting the maximum amount of available solar power.

Insufficient copper, nickel or iron for the construction of highly efficient metal oxide radiator surfaces (though amorphous carbon, e.g. lampblack, may be a substitute).

Insufficient circulating fluid (Na, NH3, CH4, O2, N2, Ne, He) for the efficient cooling of computers (rod-logic, semiconductor, helical-logic, superconductor) at specified operating temperatures.

Insufficient rare elements (Sb, In, Cd, Te, Hg, As, B, etc.) used as semiconductor dopants or as layers in solar cells or semiconductor lasers.

Insufficient silver or gold to build highly effective telescopes for observing or communicating with other civilizations.

So with this being said, even if a civilization had the incredible energy and will to create such structures, they would likely be heavily limited in how many they could construct simply by the lack of suitable materials. You can imagine such a civilization transporting entire planets worth of silicon towards MBs currently under construction. Maybe they have entire stars covered in starlifters for the sole purpose of getting enough material for this endeavor.
It'd be a great economy for cargo companies, let's say that.

@somebody With fusion reactors you would be able to produce heavier elements from stellar matter as a byproduct of energy production, but the process would take considerable time (not that constructing a Matrioshka Brain is quick, mind you...), and as far as I'm aware fusion of elements heavier than iron is endothermic. — Aug 8 at 5:36
Easy. The solution is to input a crapton of energy (and fission) :P (and the speed can be solved either with more energy or more factories/stars/whatever) — Aug 8 at 6:06
The node-neurons of the M-brain could be coordinated to function as starlifters that would magnetically squeeze the star so it will spew the necessary elements and then be caught by another coordination of matter-collecting node-neurons. Maybe each node-neuron can be equipped with supplementary fusion reactors for emergency transmutations. Better yet, maybe give node-neurons the ability to cannibalize each other. — Aug 8 at 13:15
The implementation laid out in this page... Layered structure such that even waste heat becomes useful. Moreover, I think graphene can suffice for all those element-specific insufficiencies. The 1999 paper is after all, a 1999 paper. — Aug 8 at 13:22

score 11 · Answer 3 · 2018-08-08 04:48:56Z

Besides a whole host of logistical concerns that AugustDay points out, there are a few hard theoretical bounds that preclude your first point about using these machines to perfectly simulate solar systems. These all basically hinge on one fact: the universe is very big, and there's lots of stuff in it. If you ever work in computational physics, you'll find a very large portion of the work comes down to figuring out how much you can simplify your model and not get a totally garbage result, because there's just too much stuff to simulate everything exactly.

To give a more concrete answer, let's forget about GUT level simulations and assume we just want to simulate plain old Newtonian mechanics perfectly, which is far simpler. Well, to do this, note that as a rough estimate, there are about $10^57$ atoms in the solar system, and for each one say 1000 bits is enough information to accurately describe it. This means that each time step will require on the order of $10^60$ bit operations. I'd say that a time step of at less than $10^-15$ seconds is necessary, as humans have measured time scales around this range. So that gives a value of at least $10^75$ bit operations per second of simulation time. That's astronomical-- even if you have a star at your disposal! And keep in mind this is a severe underestimate, since it doesn't take into account quantum mechanics, relativity, or even electromagnetic fields.

To get an idea of how impossibly large this number is, I'll introduce you to Landauer's Principle, which states that in irreversible computing (ie all computing we've done so far), any irreversible manipulation of information requires at least
$$kTln(2)$$
joules of energy. In space, the lowest we can get T is about 3K, which is the temperature of cosmic background radiation. You can make colder environments, but this requires additional energy for refrigeration, which would take away from the energy available for computation. Combining with the sun's luminosity of about $4*10^26 J/s$, we find the information theoretic limit of about
$$10^49operations/sec$$
for what our sun could produce. That's nothing compared to what we need, even assuming our computer works at the maximum possible efficiency (which none of our current computers even come close to). In fact, the computer using our sun wouldn't even be able to get through a single time step before the sun died!

And the situation gets even more bleak if we consider GUT level calculations like you say. In that case, the formulation of QFT requires us to create an extremely fine grid to which we assign several fields. Humans have probed to at least $10^-18m$, so if we want the simulation to convince them, we would need a grid with a resolution at least as fine as this. But this requires $10^54$ cells per cubic meter, so even simulating a single cubic meter would be pushing the limits of our computer.

Now, it might be possible to wave these concerns away by saying something along the lines of "perhaps they've utilized reversible computing", but really you can magic away any scientific objection by saying science has advanced to unforeseen frontiers. I just wanted to give you an idea of how difficult it would be to implement your plan according to modern understanding, even in an idealized world.

EDIT- I just realized I was thinking of years instead of seconds when I said the sun would die before a time step occurred. So in reality, the sun could perform about $10^6$ of the time steps I described. To be fair, my estimate of $10^60$ operations per time step is a criminal underestimate, since it doesn't take into account the operations needed to compare each particle to every other particle and do arithmetic. A more accurate number taking this into account would probably be around $10^120$ operations per time step.

Topological quantum computing (TQC) is reversible, but the problem with any quantum computer is in maintaining coherence at the necessary time period so, maybe the energy requirement can be drastically reduced if each node-neuron can use this. Introducing a GUT made everything more problematic. But as I was formulating the question, I thought an algorithm implementing GUT can simplify every calculation for every interaction. TQC based on GUT rules to run GUT simms can ponder on any future scenario, but not all at the same time, or not necessarily in order, only a future one to work on. — Aug 8 at 13:45

Griftor 1553 · Answer 4 · 2018-08-07 20:10:27Z

Why Should We?

There's a pretty famous story along these lines - the 1969 moon landing was heralded as a marvelously innovative part of American history, where NASA scientists wielded the terrifying and tremendous power of supercomputers to do calculations far beyond the scope of mankind's wildest dreams before. The sheer amount of computation we needed to do to put a man on the moon would have confounded all the greatest minds of every age up until the advent of the computer, when man was suddenly supplanted by machine as the best mathematician - performing thousands of calculations per second with no need to stop and rest, these mechanisms were miracles in and of themselves. Through their might, we performed space travel.

These days? I've got more computing power in my phone :
https://www.zmescience.com/research/technology/smartphone-power-compared-to-apollo-432/

Of course it seems like a good idea to send starter kits to set up computers on stars all around the universe; growth is human nature. But if we did so (even ignoring the issue of moving around the raw materials and manpower it would take to build such a thing, and the delays caused by space travel that can't go faster than light speed), within a matter of years, the technology would be so ridiculously behind the current level that it probably wouldn't even be worth bringing up to speed : everyone buys a new computer rather than replacing their floppy disk drive with a CD-ROM. And while it seems like you may be able to derive huge benefits from such a large machine, who's to say Siri won't be able to do the same thing in just a few years?

Trying to contemplate the kind of cosmic computational advancements youÃ¢Â€Â™re suggesting makes my brain hurt. +1. — Aug 7 at 21:00
"the technology would be so ridiculously behind the current level that it probably wouldn't even be worth bringing up to speed" - I don't think that will be the case for a civilization capable of building an M-brain, as we're already reaching the limits for how small and fast we can make computers. Take a look at this article I found - the difference between a 1988 computer and a 2008 computer is huge, but the 2008 computer is close to what you'd get today. — Aug 7 at 21:42
Yeah, our technology can still be far surpassed by quantum computers, but there has to be some limit to computational power density set by the rules of physics — Aug 8 at 3:30
"man was suddenly supplanted by machine as the best mathematician" - no, there's a difference between mathematics and mere calculation. As a mathematician, I must object ;-) — Aug 8 at 8:32
@somebody All that quantum computers need to be considered successful / powerful is quantum supremacy, which is simply the state of being better at solving a BQP problem than equivalently-priced classical computers. An exceptionally efficient quantum computer might be able to perform no more operations per second than a 3.5 MHz Zilog Z80, yet would still out perform all the supercomputers in the world combined if tasked with factoring a large semiprime. But it is true that, in theory at least, a quantum operation may take place more quickly than a semiconductor transition. — Aug 8 at 8:36

score 5 · Answer 5 · 2018-08-08 10:15:30Z

up vote
5
down vote

The creation of a Matrioshka Brain assumes certain computational feats are possible, namely the creation of strong A.I. that can mimic a human range of thought and emotion, we have no evidence whatsoever that such a thing is in fact possible. We won't build them if:

A. that level of A.I. is not possible or

B. we can't get sufficient computational density to overcome the light lag of trying to communicate across the solar system and/or

C. we can't get sufficient energy collection efficiency to power it.

That assumes that we want to build one, there are a raft of social, religious, and emotional reasons we may posit for humans never building a Matrioska Brain, they are to most modern people's way of thinking at least a bit creepy, we may never grow out of that attitude.

edited Aug 8 at 10:15

answered Aug 7 at 20:05

Ash

20.5k252127

2

Putting on shoes is creepy for cavemen and metallurgy is a sinful act of destruction reserved for pariahs in some tribes. We cannot assume that fear of being made obsolete and superstitious/social beliefs is still a thing in a civilization orders of magnitude more advanced than us compared to prehistorical mankind. Maybe even individuality is forgotten, the same way city states are not remotely self-sustainable today and were perfect mere centuries ago.
â€“Â Oxy
Aug 8 at 7:30

2

-1 Actually, we do know for a fact that strong AI is "possible". The evidence is me and you. However, we do not yet know how to build one, nor if building one will be practical. The fact that we do not know how e.g. cortical microcolumns operate on an algorithmic level does not mean that they do not exist.
â€“Â forest
Aug 8 at 8:17

1

@forest No we know that strong intelligence is possible, we don't know that A.I. is possible, it's the Artificial in Artificial Intelligence that we have no evidence for.
â€“Â Ash
Aug 8 at 10:13

2

@forest If brains are Turing Complete, how long until it runs Doom?
â€“Â Griftor
Aug 8 at 13:28

1

The creation of M-brains does not require a human-level intelligence. The software needed by an M-brain-making Von Neumann probe only requires simplistic approaches to precise fabrication techniques using asteroids, moons, planets, and the star itself as raw material and energy source. Any sophisticated algorithm for converting a star system into a M-brain can be employed. The M-brain-making Von Neumann probe can even be made biological, like an interstellar sperm swimming the void looking to impregnate a celestial egg and create a single self-sufficient organism that is a M-brain.
â€“Â Kyle Zabala
Aug 8 at 14:27

Â |Â
show 8 more comments

boxcartenant 1,812116 · Answer 6 · 2018-08-07 20:37:21Z

I see some answers about the physical or technological constraints on such a project, but I think you might be looking for answers strictly concerning ethical constraints. Here are my thoughts:

There are groups of people today who don't want us to put anything in space because they think we're polluting it with our earth germs, and thereby potentially destroying some life which we might not know about.

So here's the issue in short: The value of such a project is measurable in terms of the level of comfort and the rate of technological advancement we can achieve with the aid of such a device, but as of yet we are unable to measure the value of a single sentient lifeform.

Although we can measure the electrical movements of the brain and correlate them with different conscious activity, but we can't account for the consciousness any more than we can say that storm clouds are conscious by virtue of their electrical activity.

Science is limited that way, because it can only account for measurable phenomena, so even if we create a computational environment which mimics the electrical activity in the brain identically, in every measurable way, we can't say that it is actually conscious without personally experiencing it. And even if we download a person into the machine, we won't know if they are really conscious or if they are just simulating the behaviors associated with consciousness.

That is to say, the ends don't justify the means. No amount of quick scientific advancement can justify taking sentient life, because the value of sentience can't be measured in a way that enables us to compare it with the value of the technological advancements.

But the thing is, if the computer can perfectly simulate (i.e. behaves exactly the same as) an entire planet, meaning that anyone on the planet in question acts the exact same as their simulated selves, would it really matter whether the virtual one is conscious? After all, they'd both have the exact same thoughts — Aug 8 at 3:33
Also, you could just say that the drones only head to systems that cannot support life — Aug 8 at 3:34
A lesson I learned from classical simulation-themed movies is that any accurate-enough simulation is accurate enough. Whatever things happened to you there, so long as you retain your you-ness after it happened, truly happened to you anyway. Asking the question of what is real can also be answered by asking what isn't so yes, we know they are conscious IF we have an accurate model of consciousness. Big if. — Aug 8 at 13:54

Mike Nichols 7,07852662 · Answer 7 · 2018-08-18 03:46:23Z

Perhaps you are familiar with the works of the late Iain Banks? He is the creator of the Culture, a far-future, transhumanist utopia, which provides the setting for many of his books. In his exploration of the trajectory of human civilization into the distant future he tackled something akin to your question. I think his answer is a good one and rather than do you the disservice of paraphrasing his reasoning I provide an excerpt from his short essay "A Few Notes on the Culture".

The humans of the Culture, having solved all the obvious problems of
their shared pasts to be free from hunger, want, disease and the fear
of natural disaster and attack, would find it a slightly empty
existence only and merely enjoying themselves, and so need the
good-works of the Contact section to let them feel vicariously useful.
For the Culture's AIs, that need to feel useful is largely replaced by
the desire to experience, but as a drive it is no less strong. The
universe - or at least in this era, the galaxy - is waiting there,
largely unexplored (by the Culture, anyway), its physical principles
and laws quite comprehensively understood but the results of fifteen
billion years of the chaotically formative application and interaction
of those laws still far from fully mapped and evaluated.

By Godel out of Chaos, the galaxy is, in other words, an immensely,
intrinsically, and inexhaustibly interesting place; an intellectual
playground for machines that know everything except fear and what lies
hidden within the next uncharted stellar system.

This is where I think one has to ask why any AI civilization - and
probably any sophisticated culture at all - would want to spread
itself everywhere in the galaxy (or the universe, for that matter). It
would be perfectly possible to build a Von Neumann machine that would
build copies of itself and eventually, unless stopped, turn the
universe into nothing but those self-copies, but the question does
arise; why? What is the point? To put it in what we might still regard
as frivolous terms but which the Culture would have the wisdom to take
perfectly seriously, where is the fun in that?

Interest - the delight in experience, in understanding - comes from
the unknown; understanding is a process as well as a state, denoting
the shift from the unknown to the known, from the random to the
ordered... a universe where everything is already understood perfectly
and where uniformity has replaced diversity, would, I'd contend, be
anathema to any self-respecting AI.

Probably only humans find the idea of Von Neumann machines
frightening, because we half-understand - and even partially relate to
- the obsessiveness of the ethos such constructs embody. An AI would think the idea mad, ludicrous and - perhaps most damning of all -
boring.

This is not to say that the odd Von-Neumann-machine event doesn't crop
up in the galaxy every now and again (probably by accident rather than
design), but something so rampantly monomaniac is unlikely to last
long pitched against beings possessed of a more rounded wit, and which
really only want to alter the Von Neumann machine's software a bit and
make friends....

I'm familiar about that excerpt, although I don't see in there the answer to my question. — Aug 18 at 10:16

score 0 · Answer 8 · 2018-08-15 00:40:21Z

Some philosophers have argued that, when considered objectively, life creates more suffering than it does joy.

Better Never to Have Been: The Harm of Coming into Existence by David Benatar

If your future engineers subscribe to this belief, then they might decide that more simulations would create more suffering, and therefore the creation of more M-brains would be morally repugnant to them.

(Although it is hard to imagine how a society with such an outlook could survive more than one generation anyway. But perhaps the M-brain itself could reach the conclusion itself.)

(Then perhaps the M-brain could then be set to work doing things other than simulation. Perhaps it could try to find a way to end all universes, and thus end all suffering.)

(Or perhaps, a society with the technology to create M-brains, might have been able to prevent suffering before reaching this stage.)

Alternatively, they might wonder "What would have happened?" (FOMO)

Since Heisenberg reckons we cannot collect an exact snapshot of a solar system, we may never be able to simulate exactly what would have happened to that solar system in the future. The only way to know is to leave it alone to evolve on its own course.

This could also be tied to so metaphysical ideas. Perhaps they believe in a creator and they think it might have had a plan for the system which should not be disturbed. They could apply this rule selectively, only to systems they deem religiously significant.

They do believe in a creator - themselves. This civilization surely outgrew superstition and simply lived how they wanted. I agree they have a little FOMO, but with a grand unified theory deemed applicable to anything they can know? They'll definitely miss out on anything if they don't at least try to find out themselves what they don't (but can, with their GUT) know about yet. Toying on their simulations can lead to any of infinitely many outcomes, which they can know, if they build simulators. If they chose to leave stars alone, that's only one thing they can know, only with 100% certainty. — Aug 15 at 17:48

score 0 · Answer 9 · 2018-08-15 18:02:19Z

Better options for a mega-structure super-computer will almost certainly exist by the time that this is viable.

The idea behind the Matrioshka Brain is to harvest as much energy as possible with the highest efficiency possible. This is in order to power a mega structure super-computer. The computational power in this scenario is irrelevant since a Matrioshka Brain necessitates that the technology to make the super computers that it is composed of already exists.

Any super computer that could be made into solar orbit sized sphere could probably be made into a different shape, more easily and with fewer resources.

So when asking the question:

"In a scenario where a civilization must have the computing power of a
Matrioshka brain and has the available technology, why would they
elect not to use this idea?"

The only thing we have to think about is, "are there better potential energy sources?".

If we assume the civilization can create can create a Matrioshka Brain, then we can also assume that they already have access to the kind of computers that would be used in said Brain.

So if we have the computers, and just need to find a proper energy sources, are there any better ones than a star? Are there better energy sources that don't necessitate transporting quadrillions of units of delicate hardware into very specific locations? So many units that the fuel and materials alone may be greater than what exists in the current solar system?

Almost certainly a better and more practical power source will exist.

I can already think of a few and its not even the future yet. Make a super planet sized super computer that is powered either by kugelblitz (https://en.wikipedia.org/wiki/Kugelblitz_(astrophysics)), or antimatter, or some other high energy man made energy source.

With kugelbliz, you will still need to harvest its output using hawking radiation-absorbing, usable-energy generating devices. Still no different than an M-brain's function to generate usable energy from the stars' radiation. Microscopic blackholes are also so shortlived you have to feed them with something to prevent them from evaporating. Also, try comparing the energy outputs of a kugelblitz with a dyson swarm. Take also into account the cost of generating antimatter. — Aug 15 at 19:00

J Thomas 7767 · Answer 10 · 2018-08-18 16:35:45Z

If the aliens are like humans, I see two paths for this to happen.

They create the technology that can build a Matrioshka structure in 5 years or less. They get a sudden passion to do it, and they do. Assuming it works right the first try, then they have a Matrioshka structure. Otherwise it might take a long time before they try again. Depending partly on what they get instead of a functional Matrioshka Brain.

They start making a collection of little decisions that look logical to them. Each one looks like a good idea at the time. Eventually, in 10,000 years or less (or more) they look around and realize they have accidentally created a Matrioshka Brain.

At any point, people discussing one little decision might discuss where it's heading.

"We don't have enough bandwidth. People are getting unacceptable delays in their porn and political arguments."

"No problem, just put up another 1000 modules. We can easy charge enough to pay for it."

"You know, when you look at where we're heading with all this, if we keep going this way, in a thousand years we could have a Matrioshka Brain."

"That's good, let's discuss it sometime, maybe during the yearly office Christmas party. We have practical issues to deal with now."

I'm asking what could be the reason not to build M-brains, not what could make an M-brain happen. If your answer is that we can't make it happen yes thats true, but I specified that the M-brain construction will be automated by Von Neumann ships that we have tailored for that task. Thus, this doesn't answer the question. — Aug 19 at 8:28
My thought is that aliens that think like human beings are unlikely to make one, except: 1. If they can make the decision and get it carried out before they have time to change their minds. 2. They just go about their business trying to get a variety of goals that meet their temporary needs and desires, and at some point they look at what they've done and realize they have accidentally built a Matrioshka Brain. I say they don't need a reason not to do it. If they are like humans, heir default state is to not build one. — Aug 19 at 14:43
They do not need a reason not to build one, but I do and I'm pretty much human-like myself. If you're saying they have other priorities beside building M-brains but doing those priorities might actually end up creating one, maybe you should've elaborated what those other priorities might be. — Aug 19 at 18:42
If it takes longer than a human lifetime to build an N-brain, then anything people might want to do that takes less time is a more immediate priority. If it takes a lot of resources NOW to create an M-Brain LATER, then every competing use for those resources is an alternate priority. If you personally can get enough resources to build an M-Brain without needing anybody else's permission, then go ahead! Otherwise competing priorities are an issue. — Aug 20 at 3:40
A Matrioshka Brain takes all the resources of a star. If nobody has any competing goals, or if you personally own the star, then you shouldn't have any political difficulties with this little project.... — Aug 20 at 3:49

score 5 · Accepted Answer · 2018-08-22 12:37:12Z

I can think of a number of reasons why a civilisation that managed to build one Matrioshka Brain wouldn't ever build another, here are some of them:

slow-boat travel issues, the distances between stars means that sending conversion systems to other stars takes millions of years, assuming nothing goes wrong on the trip, a Matrioshka civilisation would of course simulate any action before taking the plunge. Sending out equipment packs to convert the neighbours will take millions of years and has at best a low chance of success. Knowing that their civilisation will never see such a plan come to fruition there's no reason to try.

light speed, so you convert neighbouring systems to create new Matrioshka Brains and? Without some sort of instantaneous communication you still don't have neighbours you can have a meaningful conversation with, there's not much point in expending resources on grand projects that don't actually do anything to enrich your existence.

induced isolationism, having a Matrioshka Brain, you can do anything you want to, as individuals and as a society, without ever leaving home anything. Such a society, with nothing to strive for, will stay home and vegetate, uninterested in anything in the outside universe.

economic collapse, building a Matrioshka Brain may be so expensive of time and resources as to deplete the civilisation that undertakes the project to the point where they simply can't mount another large scale project like an interstellar mission.

physical collapse, a Matrioshka Brain is potentially far more vulnerable to radiation damage, from events large and small, than any other form of civilisation ever thought up. A large event like a Gamma Ray burst, or a series of small ones like a flare swarm, could cripple a Matrioshka civilisation still in the construction phase, before it can look to send out envoys.

anthropic incident, if the universe is at all anthropic then the creation of a Matrioshka Brain is actually extremely dangerous. Since a Matrioshka Brain is super-sapient and super-sentient if the universe is weakly anthropic it could in fact wipe out the civilisation that tried to build it by thinking of a world without itself in it, or by thinking of something worse. If the universe is strongly anthropic a Matrioshka Brain simulation of a universe without it, like the one you propose, would result in said Matrioshka Brain disappearing along with the civilisation it supports.

You may have simply overlooked stuff like: This civilization has a formalized GUT. They can even use it in their simms. They are doing this to have neighbors. Judging by the customized heavens offered by M-brains, they're not just trying to enrich their existence: they're trying to enrich existence itself by making more of themselves exist. Inevitably, at least one of those people will want to be on other real star systems, and she/he can be in for the next project. Only probes will be sent out to start such projects. M-brains are the most vulnerable, but it's also the only standing option. — Aug 15 at 17:31
@KyleZabala There is no reason to believe that people will become any less self-centred and greedy because they make a shiny Matrioshka Brain, they'll wallow in their creation not try to improve the universe, that covers points 3 and 4. As to a Matrioshka Brain being the "only option" that's rubbish there's always another option, even if it's only "go extinct", civilisations that really go the distance will be the ones that choose the absolute safest development pathways, ones that lead away from Matrioshka Brains because of their obvious physical vulnerability. — Aug 15 at 17:53
@Ash I'm not sure you understand the anthropic principle if you believe that a computer running a simulation can cause itself to vanish. Do you want to elaborate more on how your proposed "anthropic incident" works? — Aug 15 at 18:59
@KyleZabala We don't know where Gamma Ray Bursts come from, we think they might come from supernovas but we've never seen one straight on to be able to get an origin fix, and we probably won't survive getting that data. A true Matrioshka Brain is a one trick pony, if you want to starlift, etc... you need other toys that compromise the Matrioshka Brain's functionality. The star at the core of a Matrioshka Brain is a bigger threat to it than any outside star and you don't want to mess with your civilisation's sole power source so there's no proven reason to mess with any star. — Aug 15 at 19:19
@Ash, no, that's not what anthropic means. The anthropic principle is based on the fact that the universe seems improbably perfectly suited to life. That claim suggests intelligent design. The anthropic principle answers that, no, it has to do with conditional probability. These conditions may be unlikely given no information but, given that we (human observers) exist, the probability is 1. It doesn't mean that "the observer" brings the universe into being. That's solipsism or quantum jiggery-pokery of some kind. — Aug 15 at 20:30

AugustDay 2194 · Answer 12 · 2018-08-07 20:18:02Z

Material Limits:
Matrioshka Brains are big. A megacomputer so large it is wrapped around a star in multiple layers. It is virtually impossible to comprehend just how big such a structure would be, if you were actually living in one. As Robert Bradbury puts it in this 1999 paper,

Matrioshka Brain construction is limited by the fundamental abundance of elements in their local region of space. Silicon may be excellent for building microprocessors and carbon (as diamond) excellent for building rod logic computers but neither of these elements is highly abundant in the universe as a whole. A major part of engineering MB will be the efficient partitioning of matter into the various components.

He goes on to describe some specific scenarios that would limit the construction of such a structure based upon the local situation:

Insufficient titanium, aluminum or carbon to build the maximum number of nanoscale components, particularly high pressure circulation systems and nanocomputers.

Insufficient aluminum or magnesium to build solar collecting apparatus capable of harvesting and redirecting the maximum amount of available solar power.

Insufficient copper, nickel or iron for the construction of highly efficient metal oxide radiator surfaces (though amorphous carbon, e.g. lampblack, may be a substitute).

Insufficient circulating fluid (Na, NH3, CH4, O2, N2, Ne, He) for the efficient cooling of computers (rod-logic, semiconductor, helical-logic, superconductor) at specified operating temperatures.

Insufficient rare elements (Sb, In, Cd, Te, Hg, As, B, etc.) used as semiconductor dopants or as layers in solar cells or semiconductor lasers.

Insufficient silver or gold to build highly effective telescopes for observing or communicating with other civilizations.

So with this being said, even if a civilization had the incredible energy and will to create such structures, they would likely be heavily limited in how many they could construct simply by the lack of suitable materials. You can imagine such a civilization transporting entire planets worth of silicon towards MBs currently under construction. Maybe they have entire stars covered in starlifters for the sole purpose of getting enough material for this endeavor.
It'd be a great economy for cargo companies, let's say that.

@somebody With fusion reactors you would be able to produce heavier elements from stellar matter as a byproduct of energy production, but the process would take considerable time (not that constructing a Matrioshka Brain is quick, mind you...), and as far as I'm aware fusion of elements heavier than iron is endothermic. — Aug 8 at 5:36
Easy. The solution is to input a crapton of energy (and fission) :P (and the speed can be solved either with more energy or more factories/stars/whatever) — Aug 8 at 6:06
The node-neurons of the M-brain could be coordinated to function as starlifters that would magnetically squeeze the star so it will spew the necessary elements and then be caught by another coordination of matter-collecting node-neurons. Maybe each node-neuron can be equipped with supplementary fusion reactors for emergency transmutations. Better yet, maybe give node-neurons the ability to cannibalize each other. — Aug 8 at 13:15
The implementation laid out in this page... Layered structure such that even waste heat becomes useful. Moreover, I think graphene can suffice for all those element-specific insufficiencies. The 1999 paper is after all, a 1999 paper. — Aug 8 at 13:22

score 11 · Answer 13 · 2018-08-08 04:48:56Z

Besides a whole host of logistical concerns that AugustDay points out, there are a few hard theoretical bounds that preclude your first point about using these machines to perfectly simulate solar systems. These all basically hinge on one fact: the universe is very big, and there's lots of stuff in it. If you ever work in computational physics, you'll find a very large portion of the work comes down to figuring out how much you can simplify your model and not get a totally garbage result, because there's just too much stuff to simulate everything exactly.

To give a more concrete answer, let's forget about GUT level simulations and assume we just want to simulate plain old Newtonian mechanics perfectly, which is far simpler. Well, to do this, note that as a rough estimate, there are about $10^57$ atoms in the solar system, and for each one say 1000 bits is enough information to accurately describe it. This means that each time step will require on the order of $10^60$ bit operations. I'd say that a time step of at less than $10^-15$ seconds is necessary, as humans have measured time scales around this range. So that gives a value of at least $10^75$ bit operations per second of simulation time. That's astronomical-- even if you have a star at your disposal! And keep in mind this is a severe underestimate, since it doesn't take into account quantum mechanics, relativity, or even electromagnetic fields.

To get an idea of how impossibly large this number is, I'll introduce you to Landauer's Principle, which states that in irreversible computing (ie all computing we've done so far), any irreversible manipulation of information requires at least
$$kTln(2)$$
joules of energy. In space, the lowest we can get T is about 3K, which is the temperature of cosmic background radiation. You can make colder environments, but this requires additional energy for refrigeration, which would take away from the energy available for computation. Combining with the sun's luminosity of about $4*10^26 J/s$, we find the information theoretic limit of about
$$10^49operations/sec$$
for what our sun could produce. That's nothing compared to what we need, even assuming our computer works at the maximum possible efficiency (which none of our current computers even come close to). In fact, the computer using our sun wouldn't even be able to get through a single time step before the sun died!

And the situation gets even more bleak if we consider GUT level calculations like you say. In that case, the formulation of QFT requires us to create an extremely fine grid to which we assign several fields. Humans have probed to at least $10^-18m$, so if we want the simulation to convince them, we would need a grid with a resolution at least as fine as this. But this requires $10^54$ cells per cubic meter, so even simulating a single cubic meter would be pushing the limits of our computer.

Now, it might be possible to wave these concerns away by saying something along the lines of "perhaps they've utilized reversible computing", but really you can magic away any scientific objection by saying science has advanced to unforeseen frontiers. I just wanted to give you an idea of how difficult it would be to implement your plan according to modern understanding, even in an idealized world.

EDIT- I just realized I was thinking of years instead of seconds when I said the sun would die before a time step occurred. So in reality, the sun could perform about $10^6$ of the time steps I described. To be fair, my estimate of $10^60$ operations per time step is a criminal underestimate, since it doesn't take into account the operations needed to compare each particle to every other particle and do arithmetic. A more accurate number taking this into account would probably be around $10^120$ operations per time step.

Topological quantum computing (TQC) is reversible, but the problem with any quantum computer is in maintaining coherence at the necessary time period so, maybe the energy requirement can be drastically reduced if each node-neuron can use this. Introducing a GUT made everything more problematic. But as I was formulating the question, I thought an algorithm implementing GUT can simplify every calculation for every interaction. TQC based on GUT rules to run GUT simms can ponder on any future scenario, but not all at the same time, or not necessarily in order, only a future one to work on. — Aug 8 at 13:45

Griftor 1553 · Answer 14 · 2018-08-07 20:10:27Z

Why Should We?

There's a pretty famous story along these lines - the 1969 moon landing was heralded as a marvelously innovative part of American history, where NASA scientists wielded the terrifying and tremendous power of supercomputers to do calculations far beyond the scope of mankind's wildest dreams before. The sheer amount of computation we needed to do to put a man on the moon would have confounded all the greatest minds of every age up until the advent of the computer, when man was suddenly supplanted by machine as the best mathematician - performing thousands of calculations per second with no need to stop and rest, these mechanisms were miracles in and of themselves. Through their might, we performed space travel.

These days? I've got more computing power in my phone :
https://www.zmescience.com/research/technology/smartphone-power-compared-to-apollo-432/

Of course it seems like a good idea to send starter kits to set up computers on stars all around the universe; growth is human nature. But if we did so (even ignoring the issue of moving around the raw materials and manpower it would take to build such a thing, and the delays caused by space travel that can't go faster than light speed), within a matter of years, the technology would be so ridiculously behind the current level that it probably wouldn't even be worth bringing up to speed : everyone buys a new computer rather than replacing their floppy disk drive with a CD-ROM. And while it seems like you may be able to derive huge benefits from such a large machine, who's to say Siri won't be able to do the same thing in just a few years?

Trying to contemplate the kind of cosmic computational advancements youÃ¢Â€Â™re suggesting makes my brain hurt. +1. — Aug 7 at 21:00
"the technology would be so ridiculously behind the current level that it probably wouldn't even be worth bringing up to speed" - I don't think that will be the case for a civilization capable of building an M-brain, as we're already reaching the limits for how small and fast we can make computers. Take a look at this article I found - the difference between a 1988 computer and a 2008 computer is huge, but the 2008 computer is close to what you'd get today. — Aug 7 at 21:42
Yeah, our technology can still be far surpassed by quantum computers, but there has to be some limit to computational power density set by the rules of physics — Aug 8 at 3:30
"man was suddenly supplanted by machine as the best mathematician" - no, there's a difference between mathematics and mere calculation. As a mathematician, I must object ;-) — Aug 8 at 8:32
@somebody All that quantum computers need to be considered successful / powerful is quantum supremacy, which is simply the state of being better at solving a BQP problem than equivalently-priced classical computers. An exceptionally efficient quantum computer might be able to perform no more operations per second than a 3.5 MHz Zilog Z80, yet would still out perform all the supercomputers in the world combined if tasked with factoring a large semiprime. But it is true that, in theory at least, a quantum operation may take place more quickly than a semiconductor transition. — Aug 8 at 8:36

score 5 · Answer 15 · 2018-08-08 10:15:30Z

up vote
5
down vote

The creation of a Matrioshka Brain assumes certain computational feats are possible, namely the creation of strong A.I. that can mimic a human range of thought and emotion, we have no evidence whatsoever that such a thing is in fact possible. We won't build them if:

A. that level of A.I. is not possible or

B. we can't get sufficient computational density to overcome the light lag of trying to communicate across the solar system and/or

C. we can't get sufficient energy collection efficiency to power it.

That assumes that we want to build one, there are a raft of social, religious, and emotional reasons we may posit for humans never building a Matrioska Brain, they are to most modern people's way of thinking at least a bit creepy, we may never grow out of that attitude.

edited Aug 8 at 10:15

answered Aug 7 at 20:05

Ash

20.5k252127

2

Putting on shoes is creepy for cavemen and metallurgy is a sinful act of destruction reserved for pariahs in some tribes. We cannot assume that fear of being made obsolete and superstitious/social beliefs is still a thing in a civilization orders of magnitude more advanced than us compared to prehistorical mankind. Maybe even individuality is forgotten, the same way city states are not remotely self-sustainable today and were perfect mere centuries ago.
â€“Â Oxy
Aug 8 at 7:30

2

-1 Actually, we do know for a fact that strong AI is "possible". The evidence is me and you. However, we do not yet know how to build one, nor if building one will be practical. The fact that we do not know how e.g. cortical microcolumns operate on an algorithmic level does not mean that they do not exist.
â€“Â forest
Aug 8 at 8:17

1

@forest No we know that strong intelligence is possible, we don't know that A.I. is possible, it's the Artificial in Artificial Intelligence that we have no evidence for.
â€“Â Ash
Aug 8 at 10:13

2

@forest If brains are Turing Complete, how long until it runs Doom?
â€“Â Griftor
Aug 8 at 13:28

1

The creation of M-brains does not require a human-level intelligence. The software needed by an M-brain-making Von Neumann probe only requires simplistic approaches to precise fabrication techniques using asteroids, moons, planets, and the star itself as raw material and energy source. Any sophisticated algorithm for converting a star system into a M-brain can be employed. The M-brain-making Von Neumann probe can even be made biological, like an interstellar sperm swimming the void looking to impregnate a celestial egg and create a single self-sufficient organism that is a M-brain.
â€“Â Kyle Zabala
Aug 8 at 14:27

Â |Â
show 8 more comments

boxcartenant 1,812116 · Answer 16 · 2018-08-07 20:37:21Z

I see some answers about the physical or technological constraints on such a project, but I think you might be looking for answers strictly concerning ethical constraints. Here are my thoughts:

There are groups of people today who don't want us to put anything in space because they think we're polluting it with our earth germs, and thereby potentially destroying some life which we might not know about.

So here's the issue in short: The value of such a project is measurable in terms of the level of comfort and the rate of technological advancement we can achieve with the aid of such a device, but as of yet we are unable to measure the value of a single sentient lifeform.

Although we can measure the electrical movements of the brain and correlate them with different conscious activity, but we can't account for the consciousness any more than we can say that storm clouds are conscious by virtue of their electrical activity.

Science is limited that way, because it can only account for measurable phenomena, so even if we create a computational environment which mimics the electrical activity in the brain identically, in every measurable way, we can't say that it is actually conscious without personally experiencing it. And even if we download a person into the machine, we won't know if they are really conscious or if they are just simulating the behaviors associated with consciousness.

That is to say, the ends don't justify the means. No amount of quick scientific advancement can justify taking sentient life, because the value of sentience can't be measured in a way that enables us to compare it with the value of the technological advancements.

But the thing is, if the computer can perfectly simulate (i.e. behaves exactly the same as) an entire planet, meaning that anyone on the planet in question acts the exact same as their simulated selves, would it really matter whether the virtual one is conscious? After all, they'd both have the exact same thoughts — Aug 8 at 3:33
Also, you could just say that the drones only head to systems that cannot support life — Aug 8 at 3:34
A lesson I learned from classical simulation-themed movies is that any accurate-enough simulation is accurate enough. Whatever things happened to you there, so long as you retain your you-ness after it happened, truly happened to you anyway. Asking the question of what is real can also be answered by asking what isn't so yes, we know they are conscious IF we have an accurate model of consciousness. Big if. — Aug 8 at 13:54

Mike Nichols 7,07852662 · Answer 17 · 2018-08-18 03:46:23Z

Perhaps you are familiar with the works of the late Iain Banks? He is the creator of the Culture, a far-future, transhumanist utopia, which provides the setting for many of his books. In his exploration of the trajectory of human civilization into the distant future he tackled something akin to your question. I think his answer is a good one and rather than do you the disservice of paraphrasing his reasoning I provide an excerpt from his short essay "A Few Notes on the Culture".

The humans of the Culture, having solved all the obvious problems of
their shared pasts to be free from hunger, want, disease and the fear
of natural disaster and attack, would find it a slightly empty
existence only and merely enjoying themselves, and so need the
good-works of the Contact section to let them feel vicariously useful.
For the Culture's AIs, that need to feel useful is largely replaced by
the desire to experience, but as a drive it is no less strong. The
universe - or at least in this era, the galaxy - is waiting there,
largely unexplored (by the Culture, anyway), its physical principles
and laws quite comprehensively understood but the results of fifteen
billion years of the chaotically formative application and interaction
of those laws still far from fully mapped and evaluated.

By Godel out of Chaos, the galaxy is, in other words, an immensely,
intrinsically, and inexhaustibly interesting place; an intellectual
playground for machines that know everything except fear and what lies
hidden within the next uncharted stellar system.

This is where I think one has to ask why any AI civilization - and
probably any sophisticated culture at all - would want to spread
itself everywhere in the galaxy (or the universe, for that matter). It
would be perfectly possible to build a Von Neumann machine that would
build copies of itself and eventually, unless stopped, turn the
universe into nothing but those self-copies, but the question does
arise; why? What is the point? To put it in what we might still regard
as frivolous terms but which the Culture would have the wisdom to take
perfectly seriously, where is the fun in that?

Interest - the delight in experience, in understanding - comes from
the unknown; understanding is a process as well as a state, denoting
the shift from the unknown to the known, from the random to the
ordered... a universe where everything is already understood perfectly
and where uniformity has replaced diversity, would, I'd contend, be
anathema to any self-respecting AI.

Probably only humans find the idea of Von Neumann machines
frightening, because we half-understand - and even partially relate to
- the obsessiveness of the ethos such constructs embody. An AI would think the idea mad, ludicrous and - perhaps most damning of all -
boring.

This is not to say that the odd Von-Neumann-machine event doesn't crop
up in the galaxy every now and again (probably by accident rather than
design), but something so rampantly monomaniac is unlikely to last
long pitched against beings possessed of a more rounded wit, and which
really only want to alter the Von Neumann machine's software a bit and
make friends....

I'm familiar about that excerpt, although I don't see in there the answer to my question. — Aug 18 at 10:16

score 0 · Answer 18 · 2018-08-15 00:40:21Z